| THE PUTNAM WAY JOURNAL | ||||||||||||||||
|
||||||||||||||||
|
p r e v |
How to combine the best aspects of robust statistical latent "theme" generation
models of text like topic models which are mostly unsupervized with supervized
Natural Language feature generation models for better document organization and
simultaneous multi-document snippet extraction In our most recent work we tackled this exact problem by modeling documents tagged from two different perspectives It was good to see that supervized models can be bettered independently of unsupervized models while still aiding the latter for better document organization and even providing faceted document navigation  What I Try To Do My C.V. : pdf
|
|
Academic activities (+) I am an RA from Spring 2011
(+) My previous TA assignments can be found here Frequently Visited Sites
Acitivites
Swimming
Important references (+) An all time hit reference on matrix algebra for statistics can be found in Kaare Brandt Petersen's and Michael Syskind Pedersen's Matrix Cookbook.(+) Jonathan Shewchuk's painless conjugate gradient Quite a feat! Take a read at my uncle Anadish's story in Wikipedia and in his homepageSite under construction
Caution This tiny part of the internet is under constant repairs.
Slow down as you browse and check back frequently - State Law.
Disclaimer All views expressed in these subset of web pages are
my own and does not necessarily reflect the views of my
department
or
school.
|
n e x t |
||||||||||||
| Can we re-structure the shelves to accomodate un-shelved plates easily? | ||||||||||||||||
|
What I Have Been Doing Recently Applying unsupervized statistical topic models to different problems of interest Special thanks to Jordan Boyd Graber for an important hint on parameter regularization in TagLDA and its extensions Latent Dirichlet Allocation in Slow Motion I had spent some time in drafing what goes through LDA's mind when it crunches the data of counts. Journey of parametric hierarchical multinomial generative mixture models a.k.a LDA/TagLDA |
Datasets I Have Recently Used Multidomain sentiment dataset based on Amazon Reviews Find more datasets from Sandbox Yahoo! Answers Dataset A small sample from Yahoo! Answers collected for summarization - YAHOO! Answers Jul-Aug-2009 DataSetAND09 Dataset The Blog data from the US Presidential Election campaign from August to October 2008 (All data are verified by visiting the blog sites) - AND09 Obama Mccain Speeches and Reactions in Blogs Dataset |
|||||||||||||||
|
||||||||||||||||